CVE-2026-2256

Command injection vulnerability in ModelScope's ms-agent

Published Mar 2, 2026 CVSS 3.1 · 6.5 Medium EPSS 2.1% KEV · Not listed
Description
A command injection vulnerability in ModelScope's ms-agent versions v1.6.0rc1 and earlier exists, allowing an attacker to execute arbitrary operating system commands through crafted prompt-derived input.

Analysis

AI-generated

Technical impact

A command injection flaw in ModelScope's ms‑agent allows an attacker to embed arbitrary operating system commands into prompt‑derived input, enabling the execution of malicious code on the host. The weakness is classified as CWE‑77, exposing the system to serious confidentiality, integrity, and availability risks.

Affected Systems

ModelScope ms‑agent, specifically all releases up to and including v1.6.0rc1.

Risk and Exploitability

The CVSS score of 6.5 indicates moderate severity, while the EPSS score of <1% reflects a low but measurable likelihood of exploitation. The vulnerability is not listed in the CISA KEV catalog. The likely attack vector is a remote attacker sending crafted prompt content to the agent’s exposed network interface; this inference is based on the way the ms‑agent processes prompt input.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade the ms‑agent to a version that includes the command injection fix.
  • If an upgrade cannot be performed immediately, disable or tightly restrict the prompt‑execution feature in the agent’s configuration and implement strict validation to reject shell‑command patterns.
  • Apply network segmentation or firewall rules to limit inbound connections to the agent’s interface, reducing the surface for remote exploitation.
  • Enable comprehensive logging and monitoring of ms‑agent input streams to detect and respond to attempted injection attacks.

Generated by OpenCVE AI on April 21, 2026 at 23:39 UTC.

Tracking

Sign in to view the affected projects.

Advisories

Source ID Title
Github GHSA Github GHSA GHSA-4gc2-344q-r2rw MS-Agent vulnerable to Command Injection
History
Type Values Removed Values Added
First Time appeared Modelscope
Modelscope ms-agent
Vendors & Products Modelscope
Modelscope ms-agent
Type Values Removed Values Added
Weaknesses CWE-77
Metrics cvssV3_1

{'score': 6.5, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:N'}

ssvc

{'options': {'Automatable': 'yes', 'Exploitation': 'poc', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Type Values Removed Values Added
Description A command injection vulnerability in ModelScope's ms-agent versions v1.6.0rc1 and earlier exists, allowing an attacker to execute arbitrary operating system commands through crafted prompt-derived input.
Title Command injection vulnerability in ModelScope's ms-agent
References
Weaknesses
  • CWE-77

    Improper Neutralization of Special Elements used in a Command ('Command Injection')

Sources

Status
PUBLISHED
Assigner
certcc
Published
Updated
Mar 3, 2026, 8:07 PM
Reserved
Feb 9, 2026, 3:23 PM
Status
Awaiting Analysis
Published
Mar 2, 2026, 9:16 PM
Modified
Jun 17, 2026, 10:30 AM
Severity
Moderate
Public date
Mar 2, 2026, 8:09 PM